Circadian phase position of normal subjects will be assessed by measuring their plasma melatonin levels in the evening under dim light to determine the nighttime onset of melatonin production by the pineal gland (the dim light melatonin onset, or DLMO). As a marker for circadian phase position, the DLMO appears to accurately reflect the phase-shifting effects of light. These phase-shifting effects will be evaluated in normals and compared to patients suspected of having chronobiologic sleep and mood disorders. Normal subjects will be studied on repeated occasions in order to evaluate inter- and intrasubject variability, as well as time of year effects. There will also be studies to define the phase response curve in normals. Five types of patients will be studied: advanced sleep phase syndrome, delayed sleep phase syndrome, winter depression, unipolar depression and bipolar depression. We propose to test four hypotheses: 1) intraindividual variability in the DLMO can be decreased by a week of standardized light-dark conditions; 2) human circadian rhythms are regulated by bright light according to a phase response curve similar to those described for other animals; 3) bright light is more important than social cures in synchronizing human circadian rhythms; and 4) phase advanced sleep and mood disorders will respond to bright light in the evening and phase delayed disorders will respond to bright light in the morning. We propose to do seven experiments: 1) melatonin production and response to light in normal subjects; 2) effect of social cues on circadian rhythms in normal subjects; 3) effect of light pulses on circadian rhythms in normal subjects; 4) - 6) treatment of putatively chronobiologic sleep and mood disorders with appropriately timed bright light; and 7) effect of shifting sleep on circadian rhythms in normal subjects. These investigations should help us decide how to optimally treat patients with bright light and to identify patients with abnormally advanced and delayed circadian rhythms.